Originally published on Appscribed.
Appscribed is cited by AI assistants (ChatGPT, Copilot, Google AI Overviews) roughly 2,200 times a day. Instead of guessing about answer engine optimization (AEO), we analyzed 196,554 of our own citations. Here is what we found.
The gap: citations are the new traffic
Across a 91-day window, AI assistants cited us about 2,160 times a day while sending roughly 50x fewer web clicks. If you only watch your analytics dashboard, you miss where your buyers actually see you.
Key findings
- Citations, not clicks: 196,554 citations in 91 days (about 2,160/day) versus roughly 50x fewer web clicks.
- Most citation tools undercount: the same query was cited 50% of the time in the live ChatGPT app and 0 out of 30 times via an automated API on a stripped-down search stack.
- Rank is not citation: pages ranking #80 or lower in classic search were still cited.
- Citation is probabilistic: 5 of 10 repeat fires cited, scattered rather than cached.
- Comparisons win: assistants cite third-party comparison pages for evaluative claims, and vendor pages for product facts.
- Citation is local: the same question surfaces a different roster of sources depending on the searcher's location.
The dataset
| Metric | Value | Window / method |
|---|---|---|
| Total AI citations (Google + Bing) | 196,554 | 91-day census, GSC + Bing Webmaster Tools |
| Avg citations per day | ~2,160 | derived from the window |
| Web clicks vs citations | ~50x fewer | GA4, bot-filtered |
| Citation rate, live ChatGPT app | 50% | same query, repeated fires |
| Citation rate, automated API | 0 / 30 | same query, 30 runs |
| Repeat fires cited (logged-out) | 5 / 10 | clean session, not cached |
| Lowest classic rank still cited | #80+ | observed in retrieval |
Open dataset and charts: truvery.github.io/ai-citation-study
How we measured it
We triangulated across five instruments and archived every response the moment we pulled it:
- Consumer-app probes (ChatGPT, logged out): ground-truth citation rate per query and market, fired 10+ times each.
- OpenRouter API (native web): the automated-tool baseline, 30 runs saved as raw JSON.
- DataForSEO: live SERP rank, category demand, and a reverse-lookup of competitor citations.
- Bing + Google dashboards: the 91-day census across both engines (196,554 citations).
- GA4: AI-assistant referral clicks and engagement, bot-filtered.
The seven moves that earn a citation
- Write the comparison, not the brochure. Answer "what's the best X?" better than anyone, on one page.
- Put a quotable verdict in a sentence. That is the line an assistant lifts.
- Lead with original data. Real numbers get quoted disproportionately.
- Structure for extraction: question headings, a 30-to-50-word answer beneath each, tables and lists.
- Stop treating rank as the goal. You can be cited from page eight.
- Measure a rate, per engine, per market. One check is noise; fire each query 10+ times.
- Audit the app, not the API. API tools measure the wrong stack and undercount you.
Read more
- Full study, with the archived responses and screenshots: How to Get Cited by ChatGPT
- If you would rather we do it for your product: Appscribed AI Visibility
Figures are from a fixed window and drift over time; the direction of each finding is robust. Treat the precise percentages as indicative and re-measure your own category.
Top comments (0)